Structural Change in an Open Economy - canon-igs.org · Structural Change in an Open Economy Kei-Mu...
Transcript of Structural Change in an Open Economy - canon-igs.org · Structural Change in an Open Economy Kei-Mu...
Introduction Model Model Mechanisms Dynamics Conclusion
Structural Change in an Open Economy
Kei-Mu YiFederal Reserve Bank of Minneapolis
Jing ZhangUniversity of Michigan
May 28, 20121
1The views expressed here are those of the authors are are not necessarily reflective of views of the Federal
Reserve Bank of Minneapolis or the Federal Reserve System.
Introduction Model Model Mechanisms Dynamics Conclusion
Structural Change
Share of employment (16 advanced nations)
Sector 1870 1960 1987
Agriculture 0.49 0.17 0.06Services 0.24 0.44 0.63Manufacturing 0.27 0.39 0.30
Source: Maddison (1991)
Agriculture share declines over time.
Services share rises over time.
Manufacturing share first rises and then declines over time.
Introduction Model Model Mechanisms Dynamics Conclusion
Global integration, Structural Change
World’s economies increasingly interlinked via trade.
In past 30-40 years many emerging market countries haveglobalized
Manufacturing labor shares are declining in developed nations,and rising (although not permanently) in emerging marketcountries.
Trade with emerging markets has been blamed for decliningmanufacturing employment in developed countries.
In most countries, manufacturing has the highest productivitygrowth.
Introduction Model Model Mechanisms Dynamics Conclusion
U.S. and South Korea Manufacturing Employment Share
0
5
10
15
20
25
30
1870 1880 1890 1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000
Perc
ent
Year
United States
South Korea
Labor Productivity Growth (1970-2005)
Country Agriculture Manufacturing Services
U.S. 1.5% 3.8% 0.7%South Korea 4.9% 7.0% 1.7%
Introduction Model Model Mechanisms Dynamics Conclusion
South Korea’s Manufacturing Net Exports As Share of GDP
Introduction Model Model Mechanisms Dynamics Conclusion
Manufacturing Trade and Employment
Figure: Manufacturing Net Exports and Manufacturing Employment
CANFINITAPRT
ESP
IDN
IND
JPN
KOR
PER PHL
THATWN
VEN
BOL
BRA
CHL
COL
CRIMEX
MYS
‐5%
0%
5%
10%
15%
20%
25%
ufacturing
Lab
or Sha
re, 1962‐2005
, Pe
rcen
tage Points
AUS
AUTBELCAN
DENFRAGER
ISLNLDNZL NOR
ESP
SWE
CHEGBR
USA ARG
‐25%
‐20%
‐15%
‐10%
‐15% ‐10% ‐5% 0% 5% 10% 15% 20% 25% 30%
Chan
ge in
Man
u
Change in Manufacturing Net Exports/GDP, 1962‐2005, Percentage Points
Introduction Model Model Mechanisms Dynamics Conclusion
Services Employment and Trade
list = β0 + β1tradeit + β2gdppcit + γi + εit
i : country; t: period
list : services employment share
tradeit : exports+imports as a share of GDP
gdppcit : GDP per capita in 2005 international dollars
Table: Trade and Services Labor Share
tradeit gdppcit β0 R2 Obs
Fixed Effecta 0.0801 1.23e-5 0.369 0.67 379(0.0289) (1.12e-6) (0.0251)
Introduction Model Model Mechanisms Dynamics Conclusion
Key Question
What is the effect of international trade on the process ofstructural change?
Introduction Model Model Mechanisms Dynamics Conclusion
Our Approach
We develop a two-country, three-sector model with inter- andintra-sector Ricardian trade
We study the channels by which trade affects structuralchange:
Trade delinks sectoral production and sectoral expenditure:
Closed: labor share = expenditure share
Open: labor share = expenditure share + net export share
Trade allows countries to specialize, affecting net export shares
Trade changes relative prices, affecting expenditure shares
We demonstrate two ways in which open economy cangenerate hump pattern in manufacturing employment
Introduction Model Model Mechanisms Dynamics Conclusion
Our Approach
We develop a two-country, three-sector model with inter- andintra-sector Ricardian trade
We study the channels by which trade affects structuralchange:
Trade delinks sectoral production and sectoral expenditure:
Closed: labor share = expenditure share
Open: labor share = expenditure share + net export share
Trade allows countries to specialize, affecting net export shares
Trade changes relative prices, affecting expenditure shares
We demonstrate two ways in which open economy cangenerate hump pattern in manufacturing employment
Introduction Model Model Mechanisms Dynamics Conclusion
Our Approach
We develop a two-country, three-sector model with inter- andintra-sector Ricardian trade
We study the channels by which trade affects structuralchange:
Trade delinks sectoral production and sectoral expenditure:
Closed: labor share = expenditure share
Open: labor share = expenditure share + net export share
Trade allows countries to specialize, affecting net export shares
Trade changes relative prices, affecting expenditure shares
We demonstrate two ways in which open economy cangenerate hump pattern in manufacturing employment
Introduction Model Model Mechanisms Dynamics Conclusion
Our Approach
We develop a two-country, three-sector model with inter- andintra-sector Ricardian trade
We study the channels by which trade affects structuralchange:
Trade delinks sectoral production and sectoral expenditure:
Closed: labor share = expenditure share
Open: labor share = expenditure share + net export share
Trade allows countries to specialize, affecting net export shares
Trade changes relative prices, affecting expenditure shares
We demonstrate two ways in which open economy cangenerate hump pattern in manufacturing employment
Introduction Model Model Mechanisms Dynamics Conclusion
Our Approach
We develop a two-country, three-sector model with inter- andintra-sector Ricardian trade
We study the channels by which trade affects structuralchange:
Trade delinks sectoral production and sectoral expenditure:
Closed: labor share = expenditure share
Open: labor share = expenditure share + net export share
Trade allows countries to specialize, affecting net export shares
Trade changes relative prices, affecting expenditure shares
We demonstrate two ways in which open economy cangenerate hump pattern in manufacturing employment
Introduction Model Model Mechanisms Dynamics Conclusion
Related Research
Models of structural change:
Closed-economy models: Kongsamut, Rebelo and Xie (2001)and Ngai and Pissarides (2007)
Open economy models:
Matsuyama (2009): an example with Ricardian framework
Coleman (2007): model in which large emerging marketcountry integrates with rest of world
Models of Ricardian trade: Eaton and Kortum (2002)
Introduction Model Model Mechanisms Dynamics Conclusion
Two Groups of Theories of Structural Change
Non-homothetic preferences:
Engel (1895)
Kongsamut et. al (2001)
Sector-biased productivity growth:
Baumol (1967)
Ngai and Pissarides (2007)
Closed economy frameworks
Introduction Model Model Mechanisms Dynamics Conclusion
Model Set Up
Two countries
Three sectors: agriculture, manufacturing, services
Agriculture and manufactured goods are tradable
Services are nontradable
One factor: labor with exogenous supply
Mobile across sectors, but immobile across countries
Productivity growth differs across sectors and countries
Free trade: based on Ricardian comparative advantage
Introduction Model Model Mechanisms Dynamics Conclusion
Technologies
Services: a single good Yist = AistList
Agriculture and manufacturing: a continuum of goods
yimt(z) = Aimt(z)Limt(z) z ∈ [0, 1]
yiat(z) = Aiat(z)Liat(z) z ∈ [0, 1]
A is distributed as Frechet: Fiqt(z) = exp(−Tiqtz−θ)
Goods are combined to yield composite goods for consumption
Introduction Model Model Mechanisms Dynamics Conclusion
Prices
Perfect competition in goods and factor markets
Service good price: Pist = witAist
Agricultural good price:
piat(z) = min
{w1t
A1at(z),
w2t
A2at(z)
}Manufacturing good price:
pimt(z) = min
{w1t
A1mt(z),
w2t
A2mt(z)
}
Introduction Model Model Mechanisms Dynamics Conclusion
Preferences
Tradable sector composite goods: elasticity of substitution η
Ciqt = (∫ 1
0 ciqt(z)η−1η dz)
ηη−1
Intratemporal utility: elasticity of substitution ε
Cit = (ωaCε−1ε
iat + ωmCε−1ε
imt + ωsCε−1ε
ist )εε−1
Intertemporal utility:∑∞
t=0βtU(Cit)
Budget constraint (period-by-period):
PitCit = PiatCiat + PimtCimt + PistCist = witLit
Introduction Model Model Mechanisms Dynamics Conclusion
Expenditure Shares
Expenditure share:
Xiqt =PiqtCiqt
witLit= ωεq
(Piqt
Pit
)1−ε
Aggregate price:
Pit =(ωεaP
1−εiat + ωεmP
1−εimt + ωεsP
1−εist
) 11−ε
Sectoral composite good price:
Piqt =
(∫ 1
0
piqt(z)η
η−1 dz
) η−1η
Introduction Model Model Mechanisms Dynamics Conclusion
Market Clearing Conditions
Labor markets:
Lit = List +∫ 1
0 Limt(z)dz +∫ 1
0 Liat(z)dz , i = {1, 2}
Services good markets:
Yist = Cist , i = {1, 2}
Agriculture goods markets:∑2i=1 yiat(z) =
∑2i=1 ciat(z) ∀z ∈ [0, 1]
Manufacturing goods markets:∑2i=1 yimt(z) =
∑2i=1 cimt(z) ∀z ∈ [0, 1]
Introduction Model Model Mechanisms Dynamics Conclusion
Open-Economy Equilibrium
A competitive equilibrium is a sequence of goods and factor prices{piat(z), pimt(z), Piat , Pimt , Pist , Pit , wit}∞t=0 and allocations{liat(z), limt(z), Liat , Limt , List , yiat(z), yimt(z), Yist , ciat(z),cimt(z), Ciat , Cimt , Cist , Cit}∞t=0 for z ∈ [0, 1] and i = 1, 2, suchthat given prices, the allocations solve the firms’ maximizationproblems and the household’s maximization problem, and satisfythe market clearing conditions.
Introduction Model Model Mechanisms Dynamics Conclusion
Closed Economy Equilibrium
Sectoral labor share = sectoral expenditure share
lqt =LqtLt
=wtLqtwtLt
=PqtCqt
PtCt= Xqt
Expenditure share:
Xqt = ωεq
(Pqt
Pt
)1−ε
Prices: Pqt = wtAqt
, where Aqt = T1/θqt /γ.
Introduction Model Model Mechanisms Dynamics Conclusion
Closed Economy Dynamics
Lqt = Xqt = (1− ε)(Pqt − (Xat Pat + Xmt Pmt + Xst Pst)
)
= (ε− 1)(Aqt − (XatAat + XmtAmt + XstAst)
)
ε = 1: no structural change
ε < 1: labor moves from the highest productivity growthsector to the lowest productivity growth sector
ε > 1: labor moves from the lowest productuviity growthsector to the highest productivity growth sector
Introduction Model Model Mechanisms Dynamics Conclusion
Key Implications of Closed Economy
Preferences play a major role in labor allocation across sectors.
Structural change does not occur if the elasticity ofsubstitution equals one.
With elasticity of substitution less than one:
The high productivity growth sectors experience decliningrelative prices, expenditure shares and labor shares.
Labor moves from the most productive sector to the leastproductive sector.
Introduction Model Model Mechanisms Dynamics Conclusion
Key Ingredients of Open Economy
Trade based on comparative advantage (Ricardian)
Assume country 1 has a comparative advantage inmanufacturing.
That is, under free trade,
A1mt
A2mt>
A1at
A2at.
Under free trade, the LOOP holds: p1qt(z) = p2qt(z).
Tradable composite good prices are equalized acrosscountries: P1qt = P2qt .
Introduction Model Model Mechanisms Dynamics Conclusion
Open Economy: Prices
Under Frechet distribution (and free trade):
Piqt =
[(witAiqt
)−θ+(
wjt
Ajqt
)−θ]− 1θ
Piqt
wit= 1
Aiqt
[1 +
(wjt
wit
Aiqt
Ajqt
)−θ]− 1θ
Services price: Pistwit
= 1Aist
Introduction Model Model Mechanisms Dynamics Conclusion
Prices in Open vs. Closed Economy
Piatwit
and Pimtwit
are lower in open economy
Pistwit
is the same
Pitwit
is lower in open economy
Welfare is higher in open economy
PistPit
rises, P1atP1t
and P2mtP2t
declines in open economy
P1mtP1at
is higher in the open economy
Introduction Model Model Mechanisms Dynamics Conclusion
Expenditure Shares in Open vs. Closed Economy
Relative prices and the elasticity of substitution play key rolein determining expenditure shares Xiqt .
With elasticity less than one, in both countries in openeconomy,
services expenditure shares are higher;
expenditure share of the sector with comparative disadvantageis lower;
expenditure share of the sector with comparative advantage isambiguous.
Introduction Model Model Mechanisms Dynamics Conclusion
Open Economy: Intra-Sector Trade
In addition to (sectoral) expenditure shares, another share matters:the share of sectoral spending that is on imports:
Share of country 1’s expenditure on sector q goods fromcountry 2 (under free trade):
π12qt =(w2t/A2qt)
−θ
(w2t/A2qt)−θ + (w1t/A1qt)−θ=
1
1 + (w1t/A1qt
w2t/A2qt)−θ
π12qt rises as w2t/A2qt decreases relative to w1t/A1qt
The rise is larger with larger θ (a low productivity dispersion)
Comparative advantage implies π12mt < π12at
Introduction Model Model Mechanisms Dynamics Conclusion
Open Economy: Inter-Sector Trade
Putting together these two shares:
Manufacturing net exports of country 1 as share of its GDP:
N1mt =π21mtX2mtw2tL2t
w1tL1t− π12mtX1mt
Comparative advantage implies N1mt > 0 and N1at < 0.
The net export ratio of the sector with comparative advantage ispositive in each country.
Introduction Model Model Mechanisms Dynamics Conclusion
Open Economy: Labor Allocation
Services labor share: list = List/Lit = Xist
Manufacturing labor share of country i:
limt =Limt
Lit= Ximt + Nimt
Direct contribution of trade: Nimt
Country 1 has a comparative advantage in manufacturing
N1mt > 0 and N2mt < 0
Indirect contribution of trade: Ximt
Similarly for agriculture labor share
Introduction Model Model Mechanisms Dynamics Conclusion
Labor Dynamics
Growth in manufacturing labor share:
l1mt =X1mt
l1mtX1mt +
N1mt
l1mtN1mt
First term: the expenditure effect
Second term: the trade or net export effect
Positive growth in manufacturing net export share contributespositively to labor share.
To focus on trade effect, consider case with elasticity ofsubstitution across sectors = 1; hence, X1mt = 0
Introduction Model Model Mechanisms Dynamics Conclusion
Productivity Growth and Hump Pattern in Manufacturing
Necessary condition for N1mt > 0: Amt > Aat , (Aqt =A1qt
A2qt)
Under free trade, manufacturing labor share equation:
l1mt = ωmπ11m
(wtL1t + L2t
wtL1t
)π11m = π21m: specialization term
Reciprocal of GDP share: country-size term
As manufacturing productivity grows, specialization termcontributes positively to manufacturing labor share, whilecountry-size term contributes negatively.
Each country buys more of its manufactured goods fromcountry 1 (e.g., South Korea).
If Aat > 0, country 1 relative wage grows, country 2 (e.g.,United States) purchasing power falls. Country 1 needs lesslabor to meet country 2 demand.
Introduction Model Model Mechanisms Dynamics Conclusion
Productivity Growth and Hump Pattern in Manufacturing
Initially, specialization term is dominant.
Eventually, country-size term dominates.
Once manufacturing becomes close to completely specialized,employment growth from specialization effect becomes small.
In limiting case, country 2 buys all its manufactured goodsfrom country 1, but country 2 has zero mass, so the globaleconomy is effectively just country 1.
Country 1 employment share declines until it equalsexpenditure share.
Introduction Model Model Mechanisms Dynamics Conclusion
Structural Change in Open Economy: Example 1
Preferencesε = 0.5 σ∗ = 1.0ωa = 1/3 ωm = 1/3 ωs = 1/3
Labor Endowment
L10 = 1 L20 = 10 L1t = L1t = 1.0
Sectoral Productivitiesθ = 4.0A1a0 = 1.0 A1m0 = 1.0 A1s0 = 1.0
A2a0 = A1a0(L20/L10)1/θ A2m0 = A1m0(L20/L10)1/θ A2s0 = A1s0(L20/L10)1/θ
A1at = 1.01 A1mt = 1.02 A1st = 1.0
A2at = 1.02 A2mt = 1.01 A2st = 1.0
Introduction Model Model Mechanisms Dynamics Conclusion
Structural Change in Country 1
Figure: Employment Shares, Closed and Open
0 20 40 60 80 1000
0.1
0.2
0.3
0.4
0.5
0.6
0.7
Exp & labor share, closedExp share, openLabor share, open
(a) Agriculture
0 20 40 60 80 1000
0.1
0.2
0.3
0.4
0.5
0.6
0.7
(b) Manufacturing
0 20 40 60 80 1000
0.1
0.2
0.3
0.4
0.5
0.6
0.7
(c) Services
Introduction Model Model Mechanisms Dynamics Conclusion
Structural Change in Country 2
Figure: Employment Shares, Closed and Open
0 20 40 60 80 1000.1
0.2
0.3
0.4
0.5
Exp & labor share, closedExp share, openLabor share, open
(a) Agriculture
0 20 40 60 80 1000.1
0.2
0.3
0.4
0.5
(b) Manufacturing
0 20 40 60 80 1000.1
0.2
0.3
0.4
0.5
(c) Services
Introduction Model Model Mechanisms Dynamics Conclusion
Import Shares
Figure: Import Shares
0 20 40 60 80 1000.7
0.8
0.9
1
1.1
AgricultureManufacturing
(a) Country 1
0 20 40 60 80 1000
0.05
0.1
0.15
0.2
0.25
(b) Country 2
Introduction Model Model Mechanisms Dynamics Conclusion
Wages, Prices, and Welfare
Figure: Wages, Prices, and Welfare
0 20 40 60 80 1000.4
0.5
0.6
0.7
0.8
0.9
1
1.1
1 Open2 Open
(a) Wages
0 20 40 60 80 1000
0.5
1
1.5
2
2.5
3
1 Open2 Open1&2 Closed
(b) Prices
0 20 40 60 80 1000.2
0.4
0.6
0.8
1
1.2
1.4
1.6
1 Open2 Open1 Closed2 Closed
(c) Welfare
Introduction Model Model Mechanisms Dynamics Conclusion
Structural Change and Trade Costs
Introduce iceberg trade costs
Prices:
Piqt =[(wit/Aiqt)
−θ + (τqtwjt/Ajqt)−θ]− 1
θ
Import shares:
πijqt =(τqtwjt/Ajqt)
−θ
(τqtwjt/Ajqt)−θ + (wit/Aiqt)−θ.
Decline in τqt affects Piqt and πijqt like increase in Ajqt
Decline in trade costs can also generate structural change,even in absence of biased sectoral productivity growth
Introduction Model Model Mechanisms Dynamics Conclusion
Trade Costs and Hump Pattern in Manufacturing
Suppose country 1 has comparative advantage inmanufacturing and is small relative to country 2.
Productivity levels are constant over time
As trade costs decline, specialization increases (manufacturingnet export surplus grows) and country 1 relative wage rises
Initially, specialization effect dominates country-size effect, somanufacturing labor share in country 1 rises
Eventually, country 1 labor used to satisfy country 2manufacturing demand declines, so manufacturing labor sharein country 1 falls
Introduction Model Model Mechanisms Dynamics Conclusion
Structural Change in Open Economy: Example 2
Preferencesε = 0.5 σ∗ = 1.0ωa = 1/3 ωm = 1/3 ωs = 1/3
Labor Endowment
L10 = 1 L20 = 10 L1t = L1t = 1.0
Sectoral Productivitiesθ = 4.0A1a0 = 1.5 A1m0 = 2.0 A1s0 = 1.0
A2a0 = 2.0(L20/L10)1/θ A2m0 = 1.5(L20/L10)1/θ A2s0 = 1.0(L20/L10)1/θ
A1at = 1.0 A1mt = 1.0 A1st = 1.0
A2at = 1.0 A2mt = 1.0 A2st = 1.0
Trade Costsτq0 = 2.5 τqt − 1 declines at 3% per period
Introduction Model Model Mechanisms Dynamics Conclusion
Structural Change in Country 1
Figure: Employment Shares, Closed and Open
0 10 20 30 400
0.1
0.2
0.3
0.4
Exp & labor share, closedExp share, open Labor share, open
(a) Agriculture
10 20 30 400
0.1
0.2
0.3
0.4
0.5
0.6
(b) Manufacturing
0 10 20 30 400
0.1
0.2
0.3
0.4
0.5
0.6
(c) Services
Introduction Model Model Mechanisms Dynamics Conclusion
Conclusion
International trade provides environment in which sectoraloutput and sectoral expenditure need not be equal
With neoclassical trade, comparative advantage interacts withglobal sectoral demand to determine patterns of expenditure,trade, production, and employment
We study structural change in an open economy with modelthat highlights these themes
Model yields rich insights and can potentially better explainpatterns in data
Extending model to include non-homothetic preferences,intermediate goods, and trade costs does not alter the mainimplications
Companion project: quantitative assessment
Introduction Model Model Mechanisms Dynamics Conclusion
Accounting: the U.S.
Year lmt Xmt Nmt Xmt + Nmt
1970 25.6% 27.9% -1.1% 26.8%
2000 14.5% 21.6% -4.8% 16.8%
Change -11.1% -6.3% -3.7% -10.0%
The direct trade effect accounts for one third of the decline inUS manufacturing labor share.
Introduction Model Model Mechanisms Dynamics Conclusion
Accounting: the U.K.
Year lmt Xmt Nmt Xmt + Nmt
1970 34.6% 31.0% 2.4% 33.4%
2000 16.8% 22.9% -7.8% 15.1%
Change -17.8% -8.1% -10.2% -18.3%
The direct trade effect accounts for more than one half of thedecline in British manufacturing labor share.